In the following, the design steps are explained up to the current stage:
An airplane requires power for keeping its altitude and supplying the on-board electronics constantly. During the day, the excess power coming from the solar module is stored inside the battery pack. If well designed, it is nowadays possible for a solar airplane to fly in a sustained manner, continuously, for several day and night cycles. Not surprisingly, this capability - or the endurance in general - is largely depending on the choice of the design parameters and environmental parameters such as season, geographic latitude, cloudiness, and many more. Below, two cases illustrate in a simplified manner the energy flows during the day: in the second one sustained flight is possible, while in the first one it is not.
Two cases: as opposed to the first one, the second one allows sustained flight. As performance measures, we take the endurance for the first case, and for the sustained case the excess time, the spare flight time after a full day-night cycle.
How to choose the most important design parameters is crucial for a solar-electric airplane to be operated within its specific mission characteristics. The optimization problem was reduced to the parameters wingspan b, aspect ratio λ and battery mass m_bat. For a detailed explanation on the conceptual design, the reader is referred to the publication Solar Airplane Conceptual Design and Performance Estimation.
In nutshell, the optimization is performed with respect to the performance measures introduced above: endurance for the case in which sustained flight is not possible, and excess time (reserve flight time after one day-night cicle) for the case in which it actually is possible. For the calculation of these measures depending on given environmental and design parameters, mass models for the following components are defined:
- propulsion group
The structural mass is largely depending on the size of the airplane. However, payload mass and its distribution on the airplane, as well as the design load cases have a huge influence as well that must be taken into account. This is probably best illustrated with the figure below, showing Solarimpulse and the Airbus A380, which are of similar size, but the structural weight differs by orders of magnitude.
Solarimpulse vs. A380: similar size, different mass. Source: http://www.aeroclub.com/aviondumois.htm
The optimization of the design parameters shows that it would be best to design a solar-electric airplane to be used in medium geographic latitudes at low altitude at a wingspan of around 10 meters. Operating such an Unmanned Aerial System would, however, not be very flexible: start and landing would require either an airport or some ramps. The optimization also shows, that if we fix the size to 3 meters of wingspan, sustained flight is still possible in the summer months at medium and high latitudes. In fact, this was also demonstrated with Sky-Sailor, the previous solar airplane at ASL.
This is why senseSoar features 3 meters of wingspan. Since it is equipped with more sensing and computation capabilities (thus heavier and more energy consuming payload), it will be more difficult to demonstrate sustained flight in Switzerland, but we are optimistic that we can do it.
The development in collaboration with Leichtwerk yielded a lightweight composite structure, embedding the solar modules into shell. With the help of the famous aerodynamicist Martin Hepperle, a new low-speed / high lift airfoil was developed whith the goal of achieving both a good power minimum performance as well as a benign stalling behaviour.
CAD rendering of senseSoar
The main specs are:
- Wingspan: 3 m
- Wing area: 0.725 m2
- Solar electric power (calculated) at AM1.0: 140 W
- Overall (calc.): 3.72 kg
- Batteries: 1.89 kg
- Sensor pod: 0.44 kg
- Solar module: 0.45 kg
- Speed (level flight, nominal, calc.): 10 m/s
The solar module consists of AzurSpace S32 cells assembled and encapsulated by Gochermann Solar Technology.
The battery packs consists of readily available Li-ion cells. They are accommodated inside the wing in order to distribute the weight and thus save structural mass.
The controller will be implemented on the custom ASL coreboard that various components are connected to:
- 3 Axis Accelerometers and Gyros
- 3 Axis Magnetometer
- 5 Hole Probe for airspeed vector and altitude
- uBlox GPS module
The EKF framework uses the inertial sensors for propagation and performs updates with absolute pressure measurements, the magnetic field vector measurement and GPS.
The simulation plots demonstrate that the framework is capable of recovering the position and orientation smoothely, even in the challenging motion of constant circling.
Furthermore, the angle of attack, the slip angle, the airspeed and consequently the 3-dimensional wind vector can be estimated accurately with the additional help of the 5 hole probe.
Our real-time aerodynamics and flight mechanics simulation provides the following features:
- Nonlinear lifting line implementation
- Propellers with blade element momentum theory
- Fuselage bodies considered
- C++ library
- Matlab interface
- Generic geometry definition and validation in a GUI
- Automatic derivation of parametric model for controller design
The controller itself will focus on robustness in turbulence and actuator malfunction detection.
Communication takes place over 3 ways: for high bandwidth but rather low distances (1-3 km line of sight) we use a high power WiFi module connected to the atom mother board (see below). In terms of long-range communication, a digi radio modem operating in the 686 MHz band is used. Finally, we also integrated a Spektrum 2.4 GHz receiver for standard manual RC airplane control.
A Pointgray Chameleon with a resolution of 1.3 M with wide field of view lens is integrated inside the sensor pod in order to perform vision based navigation and mapping.
A Flir Tau 320 is installed inside the airplane connected to a custom interface board allowing to stream video via Ethernet to the Atom Board:
- Framerate: 30 Hz
- Resolution: 320 x 240 (can be upgraded to VGA)
- ROS node driver and dynamic camera parameter reconfiguration
Pixel values are directly associated with an absolute temperature - allowing dedicated people search, fire localization, as well as navigation close to the terrain during night.